SpaceX Briefly Dropped Out of the $2 Trillion Club This Week. Here's My Top Buy of the 6 Stocks That Are Secure Members.

Amazon plans to spend $200B on AI infrastructure this year, a key driver of investor concern around hyperscaler capex costs but also a catalyst for AWS growth. AWS net sales rose to $37.58B in Q1 (+28% YoY), and Amazon’s custom chip business is scaling with an annual run-rate exceeding $20B. CEO Andy Jassy reiterated confidence in monetizing the 2026 AWS capex via existing customer commitments and targeting compelling operating margins/ROIC, supporting a mildly positive medium-term view despite near-term volatility.

Analysis

AMZN is the cleaner expression of the AI buildout because the market is still discounting capex as if it were pure expense, while the more important variable is monetization velocity. If the company is already pre-selling a meaningful chunk of capacity, the capex burden should compress reported free cash flow before it expands operating leverage; that usually creates a 1-2 quarter multiple drag, but it is not the same as value destruction. The key question for the next 3-6 months is whether AWS growth stays in the high-20s while capex intensity peaks; if yes, the stock can re-rate even before margins fully inflect.

Second-order winners are the infrastructure enablers that sit one layer below the headline chip race: semis, networking, power management, cooling, and data-center real estate. The more interesting loser is not AWS itself but the marginal economics of third-party AI silicon suppliers, because hyperscaler custom chips compress the addressable pool for off-the-shelf accelerators over 6-18 months. That does not negate NVDA near term—hyperscaler spend still feeds demand—but it does cap the long-duration multiple if investors start underwriting a slower mix shift toward proprietary silicon.

Contrarian view: the market is likely overfocusing on capex optics and underfocusing on unit economics inside AWS. The real falsifier is not "high spend," it is weak booking growth or evidence that installed capacity sits idle; watch for AWS growth slipping below ~20% or a second straight quarter of margin deterioration. Separately, SPCX looks like a technical rather than fundamental trade: fresh-IPO volatility, index-inclusion flows, and eventual lockup pressure can swamp business reality for weeks, so chasing the early move is high-friction unless there is a clearer earnings track record.

Key Decisions for Investors

• Buy AMZN on 2-4% pullbacks over the next 2-6 weeks; target 8-12% upside over 3 months if AWS growth stays >25% and capex is shown to be pre-committed rather than speculative. Falsify if AWS growth drops below 20% or segment margins compress for two consecutive quarters.
• Keep NVDA as a tactical long, but reduce conviction into 6-18 months and hedge with a smaller short in CBRS or a broader AI-hardware basket if custom silicon adoption becomes the dominant AWS narrative. Near-term upside remains intact from capex spend, but the long-duration multiple is at risk if hyperscaler silicon share rises faster than expected.
• Avoid chasing SPCX after the IPO pop; if you want exposure, wait for lockup-expiry weakness or a post-index-inclusion base. This is a 1-3 month technical trade, not a fundamental one, and the risk/reward is poor without a cheaper entry.
• Watch AMZN earnings/guide as the catalyst pair: if capex is maintained while AWS growth re-accelerates, the stock can outperform the Nasdaq by 500-700 bps over the next quarter; if management softens the 2026 monetization message, trim quickly because the market will punish the multiple before the numbers roll over.